Liver failure is a serious problem that effects thousands of people in the United States each year. A new form of therapy, the bioartificial liver (BAL), is in development to provide liver function to patients with liver failure prior to transplantation or until recovery of the native liver. Results of preliminary trials of a BAL containing pig hepatocytes or human C3A cells have been encouraging, but improvement is still necessary. Cell death, for example, occurs during perfusion of the BAL and limits device function and the duration of therapy. The mechanism of hepatocyte death is poorly understood, and may result from immune-mediated injury. According to this hypothesis, immune activation during a first BAL exposure may cause an accelerated response during subsequent BAL exposures. Though the BAL contains a membrane to block contact of the patient's circulation and non-autologous hepatocytes in the device, pores in the membrane allow the release of antigenic material from the BAL and the entrance of molecular mediators of rejection from the patient. The following three hypotheses will be tested as specific aims of this study: (1) the death of hepatocytes in the BAL occurs by an immune-mediated mechanism; (2) The immune response of recipients to BAL therapy is increased during secondary exposures; (3) The immune response of the recipient adversely effects the functionality of the BAL. The immune response to the BAL will be characterized by changes in antibody titers, cellular activation, and cytokine expression in the recipient. The mechanism of cell death in the BAL will be determined by histological and biochemical examination of hepatocytes after BAL therapy. Hepatocyte viability and deposition of recipient proteins in the BAL will be measured and used to assess the effects of immune response in the BAL. The BAL will be tested in healthy dogs to provide a normal, primary immune response and to allow a second BAL treatment three weeks after the first treatment. Hollow fiber membranes with mean pore diameter of 200 nanometer and 5 nanometer (approximately 100 kD molecular weight cut-off) will be compared, since these pore sizes are relevant to current clinical trials of the BAL. A better understanding of the immune response in recipients will improve BAL therapy.